Social centrality differentiation shapes the dynamics of linguistic diffusion

The diffusion of linguistic innovations depends on how speakers are embedded in their social networks. While central individuals are often thought to facilitate spread, their precise contribution remains uncertain. This study examines how the number of social ties interacts with communication scenarios to shape diffusion outcomes within a simplified computational model. Using agent-based simulations and an equation-based model, we evaluate how network position predicts the success rate, timing, and trajectory of innovation spread. Within this model, results indicate that, when individuals with more connections are assigned as the initial sources of innovations, the innovations are less likely to become successful and tend to show lower diffusion intensity and lower final diffusion levels. Under the modeled interaction rules, one-to-one interactions appear to promote stable adoption, whereas one-to-many settings favor rapid dissemination. These findings identify mechanisms and conditions under which these diffusion patterns emerge in the model: peripheral positions may provide more favorable starting points for successful diffusion, central adopters may encounter innovations earlier, and the distribution of social ties is a key factor linking both processes. They should therefore be interpreted as model-contingent results rather than general conclusions about linguistic innovation behavior.